Prometheus7 Research Institute circulated a roadmap document this week outlining the full dimensional sequence of its substrate-paradigm architecture — from the 5D primitive validated in May to a tenth-dimensional construct the institute calls the universal-unbinder, identified as the architecture's resolution point.
The document, dated to corpus ingestion on 25 May 2026, describes a twelve-rung compositional ladder. Each rung adds a new kind of operation rather than growing model size within a fixed structure. The institute reports that wall-clock training time per generation has remained in the seven-to-eleven-hour band on its research hardware, and that the cost of opening a new dimensional layer is roughly equivalent to the cost of one generation's training run.
The current operational position, as of the document's publication, is the sixth-dimensional primitive: a router-over-callables architecture that directs hidden state toward one of a set of small neural sub-modules. That primitive was in its third training attempt as of mid-May after two informative failures. The document does not report a final validation result for 6D; as of Friday it remains described as validating.
The fifth-dimensional primitive — the substrate routing manifold underlying what the institute calls the Tree of Life model family — was validated on 16 May 2026 at the 125M-parameter scale. That validation is the last confirmed step on the ladder.
The roadmap schedules the seventh through ninth primitives across the weeks immediately following 6D launch. The seventh-dimensional primitive routes across sets of callables rather than individual ones, enabling what the institute describes as parallel compositional reasoning — a coalition of specialists per token rather than one. The eighth routes across grammars, the ninth across worlds. Each step carries an explicit falsification condition: if the new primitive collapses to behavior already expressible by the prior one, the ladder is said to have reached a plateau at that point.
The tenth primitive is the document's centerpiece. The institute maps it to the universal Turing machine in computability, the holographic principle in physics, the category of all categories in mathematics, and the Kolmogorov-minimal description in information theory. The claim is that at this dimensional level the substrate becomes capable of reaching any specific from any other specific via a single unbind operation. The target date is August–September 2026.
The eleventh and twelfth primitives — a space of universal objects and a relating principle that closes the ladder back to its bottom — are classified as research dimensions. The document states plainly that they require a small research community over years and cannot be completed by one person.
What the document does not contain: performance benchmarks, external validation, or confirmed results for the 6D training run still in progress as of its writing. The roadmap is a schedule and a theoretical specification. Whether the architecture empirically clears each rung remains, as of this dateline, an open question for each step beyond 5D.